U.S. girls face significant and distinct challenges related to physical activity (PA) and subsequent health problems. Moreover, there is evidence that certain minority groups (e.g., African American and Latino children) are at particular risk for physical inactivity. Yet few studies have focused on the etiology of PA among African American and Latino girls. The proposed study will document change in PA over time among African American, Latino, and White girls, from a social contextual developmental perspective. Individual, demographic, family and home, peer, school, and neighborhood influences will be assessed over time. A total of 360 girls (N = 120 Latino, N = 120 African American, and N = 120 White girls) will be assessed annually for 4 years as part of a cohort-sequential design. Participants will be drawn from one of three age cohorts (10, 12, and 14 years of age), which, through use of the cohort-sequential design, will enable us to examine growth across an 8-year period (ages 10 to 17) with only 4 years of assessment. Girls will be recruited from 40 neighborhoods. Data will be collected via in-home assessments from the child, from a primary caregiver/parent with whom the child resides, and from a best friend of the target girl. Information will be collected via questionnaires/interviews, activity records, and accelerometers. Neighborhood GIS data, as well as public record neighborhood data, will be included as measures of neighborhood influences. The proposed study incorporates a high standard of methodological rigor, utilizing a cohort-sequential design, random recruitment, validated instruments, multiple methods of data collection, multiple informants, and state-of-the-art analytic strategies (e.g., latent growth modeling and multilevel modeling) to address the research questions and account for autocorrelations and nesting effects. Results of this study will greatly improve our knowledge of the complex individual, family, school and neighborhood mechanisms underlying PA across Latino, African American, and White girls. The ultimate goal of this research is to enhance understanding and identify malleable personal and contextual factors that can be targeted for interventions to increase girls' PA at different stages of development, across multiple social contexts, and among three ethnic groups (African American, White, and Latino).